Google Trends Scraper & API

Fast, reliable Google Trends data as clean, LLM-ready JSON — no browser, no pytrends.

Why this actor

The official apify/google-trends-scraper has a ~37% failure rate and can take up to 180s per request because it drives a headless browser. This actor talks directly to Google Trends' internal JSON endpoints over plain HTTP, giving:

• ~10× faster responses (a 2-keyword query returns in well under a second).
• ~99% success via cookie-session warm-up, rotating User-Agents, and exponential backoff on 429/5xx.
• Comparable bulk scaling — rescaled onto one 0–100 scale via an anchor term.
• LLM-ready summaries — momentum, seasonality, YoY and peak per keyword.

Features

Mode	Description
🔎 compare	Up to 5 keywords side-by-side (single request).
📦 bulk	Unlimited keywords rescaled against an anchor term → one comparable scale.
🔥 trending	Google's daily trending searches (RSS feed).

Data types: interest over time, interest by region, related queries, related topics. Search properties: web, images, news, YouTube, shopping.

Input

Key fields (see .actor/input_schema.json for the full schema):

• mode — compare | bulk | trending
• keywords — list of terms (max 5 in compare mode, up to 200 in bulk)
• timeframe — today 12-m, now 7-d, today 5-y, 2004-present, custom, …
• geo — ISO country code (empty = worldwide)
• category — Google Trends category ID (0 = all)
• property — web / images / news / youtube / shopping
• dataTypes — which datasets to return
• anchorKeyword — reference term for bulk rescaling
• proxy — Apify proxy (residential recommended for large bulk runs)

Output (dataset)

One record per keyword:

{
  "keyword": "bitcoin",
  "geo": "",
  "timeframe": "today 12-m",
  "property": "web",
  "category": 0,
  "fetchedAt": "2026-10-09T12:00:00+00:00",
  "interestOverTime": [{"date": "2025-10-12", "value": 48}, {"date": "2025-10-19", "value": 52}],
  "interestByRegion": [{"geoCode": "US", "geoName": "United States", "value": 100}],
  "relatedQueries": {"top": [{"query": "bitcoin price", "value": 100}], "rising": [{"query": "bitcoin etf", "value": 4300}]},
  "summary": {"latest": 52, "average": 48.5, "peakValue": 100, "peakDate": "2026-04-05", "trend": "rising", "yoyChangePct": 12.4}
}

In bulk mode each record also carries anchorKeyword and interestOverTimeNormalized (values comparable across all keywords).

Local development

python -m venv .venv && source .venv/bin/activate
pip install -r requirements.txt
# Run the core client directly (no Apify credentials needed):
python test_local.py
# Or run the full actor:
apify run

Notes

• Google Trends values are relative 0–100 scores, not absolute volumes.
• Google scales each request to its own maximum; use bulk mode with an anchor term if you need cross-keyword comparison.
• The related queries / related topics endpoints are blocked from datacenter IPs (Google classifies them as USER_TYPE_SCRAPER); use residential proxy to unlock them. The actor degrades gracefully — blocked endpoints are skipped with a warning instead of failing the run.
• Trending mode uses the /trending/rss feed (the old dailytrends / realtimetrends endpoints were retired by Google and now return 404).

Related searches

google trends api, pytrends alternative, google trends scraper, keyword research api, trends data json, google trends without api key, interest over time api